Molecular Ecology (2010) 19, 1312–1323 doi: 10.1111/j.1365-294X.2010.04565.x
Modern and ancient DNA reveal recent partialreplacement of caribou in the southwest Yukon
TYLER S. KUHN,*† KERI A. MCFARLANE,‡ PAMELA GROVES,§ ARNE Ø. MOOERS* and
BETH SHAPIRO†–
*Biological Sciences, Simon Fraser University, 8888 University Drive, Burnaby, British Columbia, Canada V5A 1S6, †Henry
Wellcome Ancient Biomolecules Centre, Department of Zoology, University of Oxford, South Parks Road, Oxford OX1 3PS,
UK, ‡The King’s University College, 9125-50 St. NW, Edmonton, Alberta, Canada T6B 2H3, §Institute of Arctic Biology,
University of Alaska Fairbanks, PO Box 757000, Fairbanks, AK 99775, USA, –Department of Biology, The Pennsylvania State
University, 208 Mueller Laboratory, University Park, PA 16802, USA
Corresponde
E-mail: amoo
Abstract
The long-term persistence of forest-dwelling caribou (Rangifer tarandus caribou) will
probably be determined by management and conservation decisions. Understanding the
evolutionary relationships between modern caribou herds, and how these relationships
have changed through time will provide key information for the design of appropriate
management strategies. To explore these relationships, we amplified microsatellite and
mitochondrial markers from modern caribou from across the Southern Yukon, Canada, as
well as mitochondrial DNA from Holocene specimens recovered from alpine ice patches
in the same region. Our analyses identify a genetically distinct group of caribou
composed of herds from the Southern Lakes region that may warrant special manage-
ment consideration. We also identify a partial genetic replacement event occurring
1000 years before present, coincident with the deposition of the White River tephra and
the Medieval Warm Period. These results suggest that, in the face of increasing
anthropogenic pressures and climate variability, maintaining the ability of caribou herds
to expand in numbers and range may be more important than protecting the survival of
any individual, isolated sedentary forest-dwelling herd.
Keywords: ancient DNA, caribou, conservation genetics, microsatellite
Received 2 July 2008; revision received 2 January 2010; accepted 6 January 2010
Introduction
Conservation and management are often focused on
species or populations that have small population sizes
or have been significantly impacted by recent events,
both of which are known to produce biases in estimates
of modern genetic diversity and its inferred ancestry
(Leonard 2008). Utilizing the historical perspective
afforded by ancient DNA (aDNA) can contribute signi-
ficantly to the accurate characterization of historic pop-
ulation sizes, levels of gene flow and interpopulation
relationships (Ramakrishran et al. 2005) – contributions
which in turn are critical to appropriate conservation
nce: Arne Ø. Mooers, Fax: (778) 782 3496;
and management decisions. Here, we utilize combined
modern microsatellite (short tandem repeats, STR) as
well as modern mitochondrial DNA (mtDNA) markers
and ancient mtDNA recovered from alpine ice patches
to investigate historic interpopulation relationships and
levels of gene flow for caribou (Rangifer tarandus) in
northern Canada.
Caribou (R. tarandus) are an iconic holarctic species,
representing one of the last remnants of the Beringian
megafauna (Kurten & Anderson 1980; Anderson 1984),
and the last remaining large-scale ungulate migration in
the northern hemisphere (Berger 2004). The continued
presence of caribou throughout their range is under
threat; Vors & Boyce (2009) describe a holarctic syn-
chronous population decline of nearly 60% from recent
observed population maxima in 34 of 43 major caribou
� 2010 Blackwell Publishing Ltd
ADNA REVEALS RECENT PARTIAL REPLACEMENT 1 31 3
populations for which data are available. The authors
attribute this unprecedented synchronous decline to
anthropogenic habitat modification and climate warm-
ing, demonstrating the fragility of many of these large
caribou populations. In North America, the outlook for
caribou is grim, in particular for the forest-dwelling
woodland caribou, R. tarandus caribou (Webber & Flani-
gan 1997; James & Stuart-Smith 2000; COSEWIC 2002;
McLoughlin et al. 2003; Schaefer 2003; James et al. 2004;
Wittmer et al. 2005; Hebblewhite et al. 2007; Post &
Forchammer 2008; McDevitt et al. 2009).
Forest-dwelling woodland caribou, hereafter referred
to as woodland caribou, are a paraphyletic subspecies
of R. tarandus. They are almost exclusively found in
Canada, as local extirpations have all but removed
them from the southern extreme of their range in the
United States. In Canada, woodland caribou are fur-
ther classified into one of four federally recognized
and at-risk ecotypes: Atlantic, boreal, southern moun-
tain and northern mountain (Thomas & Gray 2002).
Northern and southern mountain caribou are distin-
guished based on their winter diets (terrestrial vs.
arboreal lichens, respectively; Bergerud 2000; Thomas
& Gray 2002). The relict Atlantic caribou populations
are listed as endangered by the Committee on the
Status of Endangered Wildlife in Canada (COSEWIC
2002), whereas both southern mountain and boreal car-
ibou populations are listed as threatened. Northern
mountain caribou are classified as being of special con-
cern (COSEWIC 2002), with many populations appear-
ing to be stable or increasing (e.g., Farnell et al. 1998,
2004; COSEWIC 2002; Farnell & Gardner 2003). North-
ern mountain herds exist as small, isolated, sedentary
populations. Within our study region in the southern
Yukon, northern mountain herds range in size from
235 to 3750 individuals for the Kluane and Tay River
herds, respectively (Yukon Environment Geomatics
2008), with the majority of herds within our study
region having populations of less than 1500 (10 of 16
herds, Yukon Environment Geomatics 2008). These
small population sizes coupled with the remoteness of
their habitat, and several decades of active manage-
ment (Farnell & Russell 1984; Farnell & MacDonald
1988, 1989; Southern Lakes caribou recovery program
1996; Farnell et al. 1998; Boertje & Gardner 2000; Tho-
mas & Gray 2002; Farnell & Gardner 2003) probably
underlie the observed stabilizing trend; many herds
were in decline prior to management intervention.
Notwithstanding current trends, recent work has sug-
gested that these isolated northern mountain popula-
tions may be particularly vulnerable to the rapidly
changing arctic climate (Webber & Flanigan 1997;
Schaefer 2003; Weladji et al. 2003; Post & Forchammer
2008; Vors & Boyce 2009).
� 2010 Blackwell Publishing Ltd
Caribou conservation and management practitioners
have called for research on the temporal persistence of
caribou herds, which may provide key information to
ensure their long-term survival (Farnell et al. 2004). A
temporal perspective is also crucial for dealing with the
large natural fluctuations in population size that charac-
terize many caribou herds (Kuzyk et al. 1999; Boertje &
Gardner 2000). Here, we address these questions using
mtDNA isolated from caribou that lived in the south-
west Yukon over the last 6000 years, in combination
with eight microsatellite (STR) loci isolated from mod-
ern southern Yukon caribou herds. This combination of
genetic markers makes it possible to investigate both
short- and medium-term changes in the genetic struc-
ture and diversity of northern mountain caribou in the
southern Yukon.
Materials and methods
Sample collection
Figure 1 outlines the study area and the locations of
the relevant herds sampled for this study. In 1997,
researchers observed extremely high concentrations of
caribou faecal material preserved in the alpine Thandlat
ice patch in the southwest Yukon (Fig. 1; Kuzyk et al.
1999). Faecal material from the base of a 1.6-m ice core
was radiometrically dated at 2450 ± 50 years before
present (BP), opening up the possibility of recovering
genetic material from ancient caribou herds (Kuzyk
et al. 1999). Since 1997, 72 ice patches in the southern
Yukon have been surveyed by air and ground, with 35
yielding significant biological and archaeological
remains (Farnell et al. 2004).
We sampled 40 caribou bone and teeth samples from
13 of these ice patches for control region mtDNA
(names in italics in Fig. 1, Appendix S1, Supporting
information). These 13 ice patches are located close to
or within the present-day range of the Aishihik caribou
herd, the Southern Lakes caribou herds (Ibex, Carcross
and Atlin herds) and the Wolf Lake caribou herd.
Twenty-eight of the caribou remains were radiometri-
cally dated as part of an independent study on caribou
heavy metal contaminants. Ages for the remaining
twelve samples were estimated using the distribution of
dated caribou samples from the relevant ice patches
(G. Hare, personal communications; Appendix S1, Sup-
porting information). All dates are reported as 14C BP.
To explore the phylogeographical patterns of modern
caribou herds that overlap with the sampled ice
patches, genotypes at eight STR loci (BM4513 from
Bishop et al. 1994, and RT1, RT5, RT6, RT7, RT9, RT24
and RT27 from Wilson et al. 1997) were obtained for
395 caribou specimens from 16 herds across southern
Fig. 1 Location map of Yukon and Alaskan caribou herds showing herd ranges estimated in 2008 (adapted from Zittlau 2004; Yukon
Environment Geomatics 2008). Only herds referred to within the present study are shown: (1) Fortymile, (2) Bonnet Plume, (3) Tay
River, (4) Finlayson, (5) Chisana, (6) Kluane, (7) Klaza, (8) Tatchun, (9) Pelly, (10) Aishihik, (11) Ibex, (12) Carcross, (13) Atlin, (14)
Wolf Lake, (15) Cassiar and (16) Chase caribou herds. Colours represent region scale structure identified using 8 microsatellite loci:
(green) Cluster 1 – Southern Lakes caribou herds; (red) Cluster 2 – Aishihik herd; (blue) Cluster 3 – Fortymile, Pelly, Tay River, Wolf
Lake herds; and (yellow) Cluster 4 – Chisana and Kluane herds. Herds colour coded with blue–green strips represent herds that
were not assigned to a single cluster, but equally to Clusters 1 and 3. Inset shows detailed view of ice patch locations for samples
used in this study: (A) Gladstone (IP36), (B) Lower Gladstone (IP35), (C) East Thulsoo (IP50), (D) Vand Creek (IP34), (E) Bratnober (IP18),
(F) Upper Jo Jo (IP31), (G) Thandlat (IP1), (H) Texas Gulch (IP52), (I) Sandpiper (IP29), (J) Friday Creek (IP37), (K) Alligator (IP38), (L)
Irvine (IP80). The solid line, greyed area represents the observed extent of the eastern lobe of the White River tephra laid down
during the 1200 BP eruption. The dashed, greyed area represents the observed extent of the smaller northern lobe, deposited between
1900 and 1500 BP. Data modified from Robinson (2001).
1 31 4 T. S . KUHN ET AL.
Yukon, northern BC and eastern Alaska. These samples,
collected during the 1990s, can be used to infer recent
genetic relationships among caribou across the study
area. The herds are (refer to Fig. 1) Aishihik (n = 42),
Atlin (n = 37), Bonnet Plume (n = 19), Carcross (n = 18),
Cassiar (n = 9), Chase (n = 25), Chisana (n = 51), Finlay-
son (n = 20), Fortymile (n = 20), Ibex (n = 20), Klaza
(n = 10), Kluane (n = 21), Pelly (n = 36), Tatchun
(n = 18), Tay River (n = 20) and Wolf Lake (n = 29).
Mitochondrial DNA made use of 42 samples from
five of these modern caribou herds (Fortymile, n = 20;
Aishihik, n = 12; Ibex, n = 11 and Carcross, n = 1) for
comparison with ancient materials. These 42 samples
were chosen because they represent the only available
samples from the most likely modern relatives of the
ice patch remains (based upon the results of STR analy-
sis detailed below). Modern samples were obtained as
DNA extracts, originally isolated from caribou tissue
samples, from the Parks Canada DNA Repository
(PCDNAR), University of Alberta, Edmonton, Alberta
(Appendix S1, Supporting information). To augment
this data set, nine additional geographically appropriate
control region sequences from the Wolf Lake (n = 1)
and Fortymile (n = 8) herds were obtained from Gen-
Bank (Appendix S1, Supporting information).
Molecular methods
Modern STR extractions and amplifications were done by
Zittlau (2004). DNA was isolated from both whole blood
� 2010 Blackwell Publishing Ltd
Table 1 Comparative measures of genetic diversity
Herd
Microsatellite analysis
N A HE
Atlin 37 7.3 0.833
Carcross 18 7.4 0.791
Ibex 20 7.3 0.793
Aishihik 42 7.9 0.743
Bonnet Plume 19 9.9 0.857
Finlayson 20 9.8 0.851
Fortymile 20 10 0.866
Klaza 10 5.8 0.755
Tay River 20 8.9 0.833
Chisana 51 9.4 0.822
Kluane 21 6.4 0.748
Pelly 36 10 0.845
Tatchun 18 7.5 0.811
Wolf Lake 29 9.5 0.835
Cassiar 9 7.6 0.848
Chase 25 9.4 0.826
N, sample size; A, mean number of alleles per locus; HE,
expected heterozygosity.
ADNA REVEALS RECENT PARTIAL REPLACEMENT 1 31 5
and tissue samples. Following lysis of red blood cells
using repeated 1· ACK (0.155 M NH4Cl, 10 mM KHCO3,
1 mM EDTA, pH 7.4) washes, DNA was extracted from
white blood cells using the QIAamp Blood protocol
(QIAGEN Inc.). Twenty-five milligram tissue samples
where lysed using QIAGEN lysis buffers and proteinase
K (QIAGEN Inc.). DNA was then extracted using
QIAamp spin columns according to the DNeasy DNA
extraction protocol (QIAGEN Inc.). Polymerase chain
reaction (PCR) amplification of eight microsatellite loci
(BM4513 from Bishop et al. 1994, and RT1, RT5, RT6,
RT7, RT9, RT24 and RT27 from Wilson et al. 1997) from
both sample types was done as described in (Zittlau
2004). PCR products were separated using ABI Prism
373A and 377 Automated DNA Sequencers (PE Biosys-
terms). Genotypes were scored using GeneScan Analysis
2.0.2 and Genotyper 2.0 software (ABI).
Ancient caribou remains used in this study have been
frozen within the ice patches from the time of deposi-
tion to the time of sampling, and as such are exception-
ally well preserved, with the physical appearance and
colour very similar to modern material. This exceptional
preservation is further supported by the recovery of
well-preserved soft tissues in both faunal and archaeo-
logical remains and by the recovery of STR loci from
several ancient samples (TSK, PG and BS, unpublished
data; Farnell et al. 2004). However, as with all aDNA
research, extreme care is required to avoid contamina-
tion by modern DNA and ⁄ or the potentially complicat-
ing effect of DNA damage-induced postmortem
changes (Cooper & Poinar 2000). To minimize such
problems, we did the following. Extraction and amplifi-
cation of ancient mtDNA was carried out either in the
physically isolated aDNA facility at the Henry Well-
come Ancient Biomolecules Centre (ABC), University of
Oxford, or at a dedicated laboratory at the University of
Alaska Museum of the North (MotN) (Appendix S1,
Supporting information). These extractions where com-
pleted prior to work on any modern samples. All
extractions were analysed with appropriate extraction
and PCR negative controls inserted after every sixth
sample, all of which were sequenced. Seven samples
were separately extracted and amplified at both the
University of Oxford and University of Alaska Fair-
banks facilities, using separate primer sets, and then
compared. As a final precaution, four ancient samples,
including two samples extracted at both facilities, were
examined using bacterial cloning techniques.
Samples extracted at the ABC used a phenol–chloro-
form protocol as described in Shapiro et al. (2004).
Samples extracted at MotN used a Qiagen QIAquick
Nucleotide Removal KitTM (see Appendix S2, Support-
ing information for more detail). Multiple overlapping
primers (four pairs at ABC and five different pairs at
� 2010 Blackwell Publishing Ltd
MotN; Table S1) were used to amplify a 454-bp frag-
ment of mtDNA control region. Sequencing was per-
formed on both the forward and reverse strands using
PRISM BigDye Terminator version 3.0 (Applied Biosys-
tems) and sequenced following ethanol precipitation
(ABC) or G-50 Sephadex protocol (MotN; GE Health-
care) according to manufacturer’s instructions. To inves-
tigate sequence fidelity and DNA damage, four ancient
extracts, representing the temporal range of specimens,
were cloned using the Invitrogen (UK Ltd.) Topo-TA
bacterial cloning kit according to manufacturer’s instruc-
tions (Appendix S1, Supporting information). From
each PCR product, 10 bacterial colonies were selected
for sequencing (as above) and results compared with
those from previously recovered direct sequencing.
Genetic analysis
STR analyses. The eight STR loci are known to be in
Hardy–Weinberg Equilibrium and not to be in linkage
disequilibrium for the caribou herds used in this study
(Zittlau et al. 2000; Zittlau 2004). Comparative measures
of genetic diversity are shown for each herd (Table 1).
To examine population substructure among present-day
caribou herds across the study area, we employed a
Bayesian clustering method (Structure version 2.2; Prit-
chard et al. 2000; Falush et al. 2003) on the STR data to
estimate the number of genetic clusters that occur
within the set of sampled individuals. We used an itera-
tive approach to select the best Structure run. Each
chain was run for 500 000 iterations with a burn-in of
1 31 6 T. S . KUHN ET AL.
30 000. No prior information about an individual’s
source herd was used. We assumed an admixture
model for the ancestry of individuals and correlated
allele frequencies. We then calculated K according to
Evanno et al. (2005) to determine the highest level of
population structure.
First, as a broad-scale analysis, we examined the
whole data set of 16 caribou herds (Aishihik, Atlin,
Bonnet Plume, Carcross, Cassiar, Chase, Chisana, Fin-
layson, Fortymile, Ibex, Klaza, Kluane, Pelly, Tatchun,
Tay River and Wolf Lake) and estimated the number of
distinct genetic clusters by performing three indepen-
dent runs each for K = 1–16.
Next, we performed a finer-scale analysis by examin-
ing population substructure among the three Southern
Lakes herds (Atlin, Carcross and Ibex) and between
those herds and their neighbouring populations. To do
this, we performed two different analyses with different
sets of populations: (i) we examined the Southern Lakes
caribou and the three present-day herds that occur on
the range of ice patch samples (Aishihik, Fortymile and
Wolf Lake); and (ii) we examined Southern Lakes cari-
bou and the herds immediately adjacent to them (Pelly,
Tatchun, Tay River and Wolf Lake). For each of these
analyses, the number of distinct genetic clusters was
estimated by performing three independent runs each
for K = 1–7.
Modern and ancient mitochondrial sequences. Modern and
ancient mtDNA sequences were aligned manually in
Se-Al version 2.0a11 (Rambaut 1996). We first created a
descriptive median-joining network using Network ver-
sion 4.5 (Bandelt et al. 1999; Fluxus Engineering 2008).
We ran a Maximum Parsimony postprocessing step
(Polzin & Daneschmand 2003), and down-weighted the
highly variable mutational positions 183, 217, 268, 315
and 444 by 50%, as suggested in the Network user’s
guide. To identify maximally differentiated, but geo-
graphically homogenous population structure without
conditioning the analysis on a priori population struc-
tures (groupings of one or more populations), we
employed the program Spatial Analysis of MOlecular
Variance, SAMOVA version 1.0 (Dupanloup et al. 2002;
http://web.unife.it/progetti/genetica/Isabelle/samova.
html). The SAMOVA analysis was run using 1000 initial
states and geographic groupings from K = 2 to 6. We
selected K by identifying a plateau in the decrease of
the total genetic variance that could be explained by
differences among populations within groups. Alligator
and Sandpiper ice patches (Fig. 1; Appendix S1, Sup-
porting information) contain ancient samples spanning
a large time period (2000 years and �5500 years,
respectively) and have bimodal age distributions. We
therefore ran the SAMOVA analysis twice: first consider-
ing them a single ‘population’ and second with the two
ice patches each subdivided into ‘recent’ and ‘ancient’
samples (recent ice patch samples are denoted with an
‘R’, e.g. SandpiperR). For both populations, recent sam-
ples represent those <1000 BP, with ancient samples
>>1000 BP as identified by the bimodal age distribution.
As each population within SAMOVA must have a unique
geographic location, the coordinates of recent samples
from Alligator and Sandpiper ice patches were shifted
<0.05� north and west of the locations used for ancient
Alligator and Sandpiper ice patch samples.
Although SAMOVA is not preconditioned on population
structure (predefined groups of populations), individual
samples must be placed within individual populations
a priori. Defining these populations a priori and not
explicitly incorporating information about variation in
population range locations through time, sample age or
evolutionary history may influence the optimal SAMOVA
groupings. To properly incorporate these factors into
our investigation, we utilized Bayesian phylogenetic
reconstruction approach using the Bayesian genealogi-
cal inference package BEAST version 1.4.8 (Drummond &
Rambaut 2007). A major advantage of BEAST is that it
allows the explicit incorporation of noncontemporane-
ous DNA sequences in a coalescent-based inference
framework. Each BEAST analysis was run for 30 000 000
iterations, sampling every 3000, with the first 3 000 000
steps discarded as burn-in. We selected the optimal
evolutionary model using the software ModelGenerator
version 0.85 (available from http://bioinf.may.ie/soft-
ware/modelgenerator/). For all BEAST analyses, we used
an Hasegawa–Kishino–Yano model of nucleotide subs-
titution with gamma-distributed rate heterogeneity
(Hasegawa et al. 1985) assuming a strict molecular clock
and a constant population size. We also tested a more
generalized model allowing for dynamic population
changes (the Bayesian Skyline plot; Drummond et al.
2005), however, Bayes Factor analysis showed stronger
support for the simpler constant population size model
(results not shown). Estimated sample sizes for each
parameter in each independent chain were examined
using Tracer version 1.4 (Rambaut & Drummond 2007)
to evaluate mixing, and plots of posterior samples were
evaluated by eye to ensure each run reached stationarity.
Results
Sequence authenticity
Consistent with the level of preservation of the ancient
remains used in this study, we observed no evidence of
contamination or DNA damage. Cloned reagent and
PCR negative controls produced no recoverable DNA.
The four extracts selected for bacterial cloning produced
� 2010 Blackwell Publishing Ltd
(a)
ADNA REVEALS RECENT PARTIAL REPLACEMENT 1 31 7
identical sequences to those previously recovered from
direct sequencing of these same extracts. Finally, each
of the seven sequences that were extracted and
sequenced at independent facilities, using different
extraction and amplification protocols, resulted in iden-
tical sequences. This exceptional preservation is further
demonstrated by recovery of STR loci from several
ancient samples (TSK, PG and BS, unpublished data;
J. Haile, personal communication). However, because of
the limitations of the available ancient samples, detailed
STR analysis of these ancient samples could not be
performed.
(b)
Fig. 2 Mean proportion of each caribou herd assigned to
genetic Cluster 1, 2, 3 or 4, based on Bayesian clustering analy-
ses of individual genotypes at eight STR loci. (a) Broad-scale
analysis of the whole STR data set of 16 caribou herds; (b)
finer-scale analysis of Southern Lakes caribou and the herds
covering the range of ice patch samples.
Modern microsatellite DNA analyses
The clustering analysis of Structure revealed four
distinct genetic clusters among the 16 caribou herds:
Cluster 1 included the Southern Lakes herds (Atlin,
Carcross and Ibex); Cluster 2 included the Aishihik
herd; Cluster 3 included the Bonnet Plume, Finlayson,
Fortymile, Klaza and Tay River herds; and Cluster 4
was composed of the Chisana and Kluane herds
(Fig. 2a). Assignment probabilities were highest among
individuals belonging to Cluster 2 (Aishihik). Caribou
from Pelly, Tatchun and Wolf Lake assigned almost
equally between Clusters 1 and 3. The remaining herds
could not be strongly assigned to any one cluster.
At the finer scale, analyses examining substructure
among the Southern Lakes caribou and the present-day
herds that occupy the range of ice patch samples (Aishi-
hik, Fortymile and Wolf Lake) revealed three distinct
genetic clusters similar to those identified at the broad-
scale analysis: Cluster 1 contained the Southern Lakes
caribou; Cluster 2 corresponded to Aishihik caribou;
and Cluster 3 corresponded to Fortymile caribou
(Fig. 2b). Again, assignment probabilities were highest
for Cluster 2. Wolf Lake caribou were equally likely to
be assigned to Clusters 1 and 3, suggesting that the car-
ibou from the Wolf Lake herd may represent an admix-
ture of both Fortymile and Southern Lakes herds or be
derived from populations ancestral to both the Forty-
mile and the Southern Lakes herds.
To estimate the spatial extent of Cluster 1 (Southern
Lakes caribou), we analysed only the three Southern
Lakes caribou (Atlin, Carcross and Ibex) and immedi-
ately adjacent herds (Pelly, Tatchun, Tay River and
Wolf Lake herds; herds that did not strongly associate
with any one cluster in the regional scale analysis). At
this spatial scale, the Atlin, Carcross and Ibex herds
form a weak genetic cluster, observed when K = 2
(mean ln likelihood = )6464). However, the estimated
ln likelihood of the data was largest and most strongly
supported for K = 1 (mean ln likelihood = )5906), sug-
gesting a single population.
� 2010 Blackwell Publishing Ltd
Modern plus ancient mtDNA analyses
All ancient and modern haplotypes included herein are
representatives of the Beringian-Eurasian Lineage (as
defined by Dueck 1998; Flagstad & Røed 2003; and
McDevitt et al. 2009; analysis and results not shown).
As a result, we felt it inappropriate to include many
previously published modern haplotypes that are repre-
sentatives of the North American Lineage as the very
large genetic distance between these two groups would
probably have diminished our ability to assess regional
phylogeographic patterns.
The median-joining network shows considerable
sequence diversity within the caribou mtDNA control
region (Fig. 3). Two high-frequency haplotypes are cen-
trally located within the network (Fig. 3; haplotypes
Rt003 and Rt007), and correspond to the only two hapl-
otypes observed in the Aishihik herd. The haplotypes
Fig. 3 Median-joining network of all ancient and modern haplotypes. Appendix S1 (Supporting information) provides further
information on sampling locations, ages and haplotype frequencies (haplotype labels identify unique haplotypes only). Colors
indicate clusterings identified using eight STR loci. Clade 1, outlined in grey, is the only visually distinct group within the diverse
network, and its member haplotypes corresponds closely with haplotypes found in populations identified using Spatial Analysis of
MOlecular Variance, SAMOVA (Table 2), and results from BEAST (Fig. 4).
1 31 8 T. S . KUHN ET AL.
found in the Carcross and Ibex herds tend to be on
short branches related to these two common sequences.
In contrast, haplotypes observed in the larger Fortymile
herd (haplotypes in blue in Fig. 3) are distributed
throughout the network, but tend to be on longer
branches arising from the central portion of the net-
work. Ancient haplotypes show two similar patterns:
short branches generally arising from a central ancient
haplotype, or long branches similar to the Fortymile
haplotypes. The network identifies one small grouping
of haplotypes (Clade 1, Fig. 3) that appear closely
related, with the remaining haplotypes forming a
diverse cluster with no clear internal structures (Fig. 3).
Clade 1 is composed of the dominant haplotypes from
two of the Southern Lakes herds (Ibex and Carcross),
one of the two haplotypes observed in the Aishihik
herd and several ancient haplotypes, all more recent
than �800 BP. Clade 1 also contains a number of haplo-
types found only in the Fortymile herd. The herds that
represent haplotypes identified within Clade 1 corre-
spond to those herds indentified within Cluster 1 of
the STR analysis, with one key difference: Clade 1 also
includes several Aishihik and Fortymile herd haplo-
types. For this reason, we distinguish Clade 1 as repre-
senting mtDNA genetic structure and Cluster 1 as
representing the corresponding STR genetic structure.
The mtDNA SAMOVA results for the multiple K values
examined for mtDNA data are summarized in Table 2.
These results suggest that three clades (K = 3) account
for the optimal genetic ⁄ geographic structure, accounting
for 25.5% of variation at the Among Groups level (Fct =
0.255, P > 0.0001). The three groups are: (i) Carcross,
Ibex, SandpiperR, Upper Jo Jo; (ii) Fortymile, Aishihik,
Wolf Lake, and all Ancient samples excluding SandpiperR
and Upper Jo Jo; and (iii) Thulsoo. Ignoring the age
dichotomies in Alligator and Sandpiper ice patches pro-
duced a similar but less optimal structure (K = 3, 22.6%
of variation at the Among Groups level, Fct = 0.226,
P > 0.0001). Geographic position was found to have no
effect on the ability of SAMOVA to recover the optimal
outcome (results not shown). Excluding the Thulsoo
sample from the analysis returned an optimal K-value of
K = 2, with the same clades described above (results not
shown).
The Bayesian reconstructions, which were performed
so as to incorporate explicitly the temporal component
of the data, reveal a single well-supported clade (pos-
terior support of 0.93; Fig. 4) that includes the same
haplotypes isolated in Clade 1 identified in the network
analysis (Fig. 3). The corresponding herds which con-
tained these Clade 1 haplotypes mirror the SAMOVA
grouping of the modern Ibex and Carcross herds, plus
recent (<1000 BP) Sandpiper and Upper Jo Jo ice patch
samples.
Discussion
Our analyses reveal a dynamic recent history for cari-
bou in southern Yukon. The rapidly evolving STR loci
identify strong interherd genetic structure at broad geo-
graphic scales. More slowly evolving maternally inher-
ited mtDNA supports some of these multi-herd genetic
groupings (specifically a clade encompassing herds
from the Southern Lakes region), and when coupled
with the STR data, identify past genetic mixing and
possible founder effects (specifically the lack of correla-
tion between STR and mtDNA markers for the Aishihik
herd). Finally, the addition of data from ancient caribou
allow novel inference about the medium-term geo-
graphic and genetic stability of caribou populations,
and recover an entirely unforeseen partial replacement
event in the southern-most part of the Yukon around
1000 BP.
In a recent analysis of caribou diet, Farnell et al. (2004)
analysed faecal material collected at the Friday Creek ice
patch, and found that ice patch caribou had a diet simi-
� 2010 Blackwell Publishing Ltd
Table 2 Results from Spatial Analysis of MOlecular Variance (SAMOVA) for K = 2–6
K Fst (%var) Fsc (%var) Fct (%var) Structure
2 0.40515 (59.48)* 0.21069 (15.88)* 0.24637 (24.64)** (Thulsoo) (Remainder)
3 0.34865 (65.13)* 0.12612 (9.40)* 0.25465 (25.46)* (Thulsoo) (Carcross, Ibex, SandpiperR, Ujo) (Remainder)
4 0.34579 (65.42)* 0.10920 (8.02)* 0.26559 (26.56)* (Thulsoo) (Carcross, Ibex, SandpiperR, Ujo)
(Wolf, Irvine) (Remainder)
5 0.34419 (65.58)* 0.10023 (7.31)* 0.27113 (27.11)* (Thulsoo) (Carcross, Ibex, SandpiperR, Ujo)
(Wolf, Irvine) (Vand) (Remainder)
6 0.30317 (76.82)* )0.10242 ()7.14)* 0.30317 (30.32)* (Thulsoo, Gladstone, Bratnober) (Carcross, Ibex,
SandpiperR, Ujo) (Wolf, Irvine) (Vand, Friday,
Thandlat, AlligatorR, Texas) (Aishihik,
AlligatorA, SandpiperA) (Fortymile)
*P > 0.005; **P = 0.054. Optimal grouping is K = 3 (bold), as increased K values beyond K = 3 result in minimal increase to Fct. The
analysis was also run excluding the Thulsoo ice patch sample (not shown), which produced similar results to those reported here and
an optimal K of K = 2. SAMOVA consistently identifies the Southern Lakes caribou herds as distinct from surrounding herds at all K
values.
ADNA REVEALS RECENT PARTIAL REPLACEMENT 1 31 9
lar to the small sedentary northern mountain caribou
herds that inhabit the Southern Lakes region today (Ibex,
Carcross and Atlin herds; STR Cluster 1). This distin-
guishes these ancient caribou from the large migratory
Fortymile herd whose historic range encompassed this
ice patch as recently as 100 BP (Boertje & Gardner 2000).
This dietary analysis is consistent with modern caribou
behaviour and habitat usage, where northern mountain
caribou undergo modest altitudinal migrations from
their winter range in valley bottoms to higher elevations
during the summer. At higher elevations, these caribou
are known to aggregate on alpine ice patches to avoid
insect harassment (Weladji et al. 2003). In contrast, the
migratory Fortymile herd would have occupied its
southern range limits during the winter months, when
individuals are most dispersed, the ice patches are cov-
ered in deep winter snow and suitable habitat is only
found in valley bottoms. In support of these divergent
life histories, we find little or no gene flow between the
Fortymile herd and the herds of the Southern Lakes
region. These results are in contrast to observations from
Alaska, where an expanding migratory herd introgres-
sed with the resident sedentary herds to such an extent
that individual herd identification became impossible
(Hinkes et al. 2005). Interestingly, sedentary northern
mountain caribou herds that surround the Southern
Lakes region (Bonnet Plume, Finlayson, Klaza and Tay
River herds; members of STR Cluster 3), which are
behaviourally similar to the Southern Lakes herds, are
genetically indistinguishable, using fast evolving STRs,
from the migratory Fortymile herd. This close genetic
history between migratory and sedentary ecotypes sug-
gests that significant gene flow still exists between these
geographically dispersed herds, a result that is at odds
with present management frameworks.
� 2010 Blackwell Publishing Ltd
Short tandem repeats and mtDNA data provide dif-
fering results with regards to the evolutionary relation-
ships among the sedentary herds. In the STR analysis,
the Aishihik herd forms a distinct cluster (STR cluster
2), suggesting strong isolation from surrounding herds.
Had this herd been isolated from surrounding herds for
several 1000 years, we would have expected that Aishi-
hik caribou and the ancient caribou recovered from ice
patches within their geographic range (Gladstone, Lower
Gladstone and Thulsoo ice patches) would form a distinct
clade. However, all of the modern and ancient mtDNA
haplotypes from this herd are shared with the sur-
rounding herds. Unfortunately, we were unable to iso-
late useable STRs from these ice patch remains. The
distinctiveness of the Aishihik herd observed using STR
markers is then best explained as the result of recent
founder effects or population bottleneck coupled with
short-term isolation. Population estimates for the Aishi-
hik caribou herd document a 50% reduction in herd
size between 1981 and 1993, which precipitated an
extensive predator control and management strategy
designed to stabilize this economically and culturally
important herd (Hayes et al. 2003).
In contrast to the Aishihik herd, both STR and
mtDNA results provide evidence for the distinctiveness
of the Ibex, Carcross and Atlin herds of the Southern
Lakes region (mtDNA Clade 1; STR Cluster 1). If these
herds had been isolated from others for the last several
1000 years, caribou recovered from Alligator, Bratnober,
Friday Creek, Texas Gulch, Thandlat, Sandpiper and
Vand Creek ice patches should have fallen within this
group. However, while many recent (<1000 BP) ice
patch samples from the Southern Lakes region show
close genetic affinity with the herds present on the
landscape today, our results identify a distinct transi-
3000.0
Rt003_ACH/FCH@0
Rt014_IP37@3220
Rt003_IP36@120
Rt038_FCH@0
Rt015_IP38@1000
Rt044_GC@0
Rt009_IP36@5000
Rt027_IP38@3720
Rt001_IP36@3790
Rt006_IP36@3480
Rt031_FCH@0
Rt028_ICH@0
Rt025_IP38@360
Rt045_GC@0
Rt018_IP80@190
Rt006_IP37@2500
Rt023_IP37@2320
Rt003_IP36@5000
Rt013_IP39@2510
Rt022_IP37@3740
Rt021_IP35@4190
Rt026_IP50@190
Rt012_IP37@3890
Rt006_IP36@5000
Rt032_ICH@0
Rt010_IP36_a@5000
Rt043_GC@0
Rt006_IP18@5000
Rt012_IP29@5710
Rt003_IP54@90
Rt008_IP37@1940
Rt014_IP1@3140
Rt039_ICH@0Rt005_IP1@3000
Rt019_IP31@790
Rt024_IP37@2550
Rt011_FCH@0
Rt010_IP36_b@5000
Rt004_IP37@2500
Rt002_IP36@3820
Rt030_FCH@0
Rt020_IP34@2700
Rt029_FCH@0
Rt034_FCH@0
Rt035_CCH@0
Rt014_IP52@5000
Rt041_GC@0
Rt011_IP37@3760
Rt007_ACH/FCH@0
Rt008_IP37@2500
Rt026_FCH@0
Rt033_FCH@0
Rt014_IP37@3150
Rt016_IP52@6320
Rt018_WCH@0
Rt037_ICH@0
Rt046_GC@0
Rt040_FCH@0
Rt036_FCH@0
Rt007_IP38@2340
Rt018_IP18@4760
Rt017_IP37@2500
Rt042_GC@0
Rt006_IP36@4830
0.92
1
0.89
0.96
0.92
0.99
0.8
0.86
0.93
1
1
0.86
0.98
1
0.84
1
1
0.78
Fig. 4 Maximum Clade Credibility genealogy summarized
from the BEAST analysis described in the main text. Posterior
support values greater than 0.75 are shown adjacent to each
node. Line styles indicate structure obtained from Spatial Anal-
ysis of MOlecular Variance (SAMOVA) (Table 2). Branch lengths
are measured in radiocarbon years before present, with the rate
estimated from dated sequences. The only well-supported
clade comprises the same samples that are identified as Clade
1 in the median-joining network. Similar to the tree topology,
the optimal SAMOVA structure includes the Southern Lakes
herds as a separate group (mtDNA Clade 1), however, as the
SAMOVA approach requires samples to be assigned to an indi-
vidual population a priori, several individual haplotypes
appear to represent recent migrants to or from the Southern
Lakes mtDNA Clade 1. All but one ancient samples dated
before �1000 BP (indicated by the thin grey vertical line)
group outside the well-supported Clade 1, indicating that
ancient caribou from the Southern Lakes region do not repre-
sent the ancestral source of present-day Southern Lakes herds.
1 32 0 T. S . KUHN ET AL.
tion occurring �1000 BP. Caribou samples older than
this invariably fell outside the well-supported Clade 1,
and instead were indistinguishable from modern Aishi-
hik and Fortymile samples and ice patch samples from
the Aishihik range.
While the proposed replacement event explains most
of the differences in caribou genetic diversity observed
before and after 1000 BP, there are several exceptions.
For example, the small but diverse Ibex caribou herd
contains mtDNA haplotypes found both within and
outside Clade 1. These haplotypes are not shared with
any of the herds tested here, or any of the previously
published sequences (broad-scale analysis results not
shown). Similarly, a small number of Fortymile haplo-
types are found within Clade 1. These exceptions may
represent recent dispersal and gene flow within the
Southern Lakes region or ancestral haplotypes that sur-
vived the replacement event. Given this, we adopt a
conservative interpretation of partial replacement to
explain the 1000 BP replacement event.
The timing of the proposed partial replacement event
is consistent with two large-scale environmental distur-
bances within the southern Yukon: (i) the eruption and
deposition of the White River tephra, and (ii) the global
climate warming of the Medieval Warm Period (MWP).
The White River tephra was deposited in two lobes:
the first, smaller, northern lobe between 1900 and
1500 BP, and the much larger, eastern lobe �1200 BP
(Fig. 1; Clague et al. 1995; Robinson 2001). The eastern
lobe represents the largest known Holocene stratovol-
cano (Lerbekmo 2008), with an estimated 47 km3 of vol-
canic ash deposited across c. 540 000 km2 (Robinson
2001; Lakeman et al. 2008). Ash thicknesses of up to
30 cm having been recorded more than 300 km from
the source volcano (Bostock 1952). Perhaps unsurpris-
ingly, the deposition of this lobe has been linked to
extensive archaeological upheaval in the southern
Yukon (Workman 1979; Moodie et al. 1992; Robinson
2001; Hare et al. 2005).
The detrimental effects of ash fall on livestock and
caribou have been observed from 2.5–10 cm of ash
deposited from the 1912 Mount Katmai eruption in
Alaska (Jagger 1945) and as little as 1.9 cm of ash fol-
lowing the 1947 eruption of Hekla, in Iceland (Malde
1964). Given that much of the Southern Lakes region
falls within the 5 cm isopach (ash depths of 5 cm or
greater) of the eastern lobe of the White River tephra
(Fig. 1), it seems probably that this cataclysmic event
� 2010 Blackwell Publishing Ltd
ADNA REVEALS RECENT PARTIAL REPLACEMENT 1 32 1
had a noticeable effect on the distribution of caribou
populations living within the region at the time.
The appearance of Clade 1 in the Southern Lakes
region at �1000 BP follows a 400 year period during
which no remains were preserved within the sampled
ice patches (1440–1030 BP, Farnell et al. 2004). This hia-
tus is coincident with the increase in global temperatures
associated with the MWP (Farnell et al. 2004; Helama
et al. 2009). An increase in mean July temperatures is
also inferred from both pollen and chironomid accu-
mulation rates within our study region (Bunbury &
Gajewski 2009). Warmer temperatures mean more fre-
quent periods of winter thawing, increased winter snow-
fall, increased insect harassment and loss of perennial
snow patches (Thomas & Gray 2002; Weladji et al. 2003),
all of which have detrimental effects on caribou. It is
therefore possible, even probable, that the MWP also
played a role in the extirpation of caribou from the
Southern Lakes region, with subsequent recolonization
occurring as temperatures cooled during the Little Ice
Age. Interestingly, and as a result of warmer summer
temperatures observed in recent years within our study
region, the vast majority of sampled ice patches, which
had persisted for upwards of 7000 years, no longer exist.
Our combined analysis of ancient and modern caribou
provides several insights for caribou management and
conservation. First, combined STR and mtDNA analyses
demonstrate that the Southern Lakes caribou herds are
distinguishable from other caribou herds, but that recent
shared ancestry or gene flow has blurred this genetic
division on smaller spatial scales. Our results support
management of these three herds (Atlin, Carcross and
Ibex) as a single management unit, and provide genetic
evidence in support of active management practices
designed to ensure their long-term persistence.
Interestingly, however, the most distinctive herd based
on STR data, the sedentary Aishihik herd, is indistin-
guishable from the adjacent migratory Fortymile herd
when using mtDNA. However, a dichotomy between
STR and mtDNA analysis is not unsurprising given
observations for woodland caribou from the Central
Rocky Mountains of British Columbia and Alberta
(McDevitt et al. 2009). In contrast, where McDevitt et al.
(2009) observed a dichotomy between markers for herds
of the same ecotype with recent mixing of two divergent
mitochondrial lineages, the dichotomy observed in
southern Yukon is between different ecotypes, but where
we see isolation of a formerly uniform mitochondrial
lineage. The unique genetic information contained within
the Aishihik herd is then a very recent phenomenon.
Coupled with the observed genetic relationships between
woodland herds of the central Yukon, we suggest that
the total unique genetic diversity contained within
these herds remains very small and of recent origin. As a
� 2010 Blackwell Publishing Ltd
result, the loss of any one herd would not have a signifi-
cant effect on the overall genetic diversity of caribou
within the region. It is, however, important to acknowl-
edge the economic and cultural importance of each indi-
vidual herd within its region. Our results underscore the
importance of using multiple lines of evidence when
making appropriate management decisions, as analysis
of STR markers alone could have lead to different man-
agement priorities regarding the Aishihik herd.
Finally, by adding a temporal perspective to our anal-
ysis, we identify a recent, partial replacement event that
is not discernible from modern data alone. This replace-
ment event was most probably caused by changes in
caribou habitat as a result of either the MWP or the
deposition of the White River tephra. At present, we
cannot distinguish between these two closely timed
events, however, increased sampling and stratigraphic
coverage of the 1500–500 BP period may be sufficient to
identify a single causal mechanism. Whether this
replacement event was caused by a large volcanic erup-
tion or increased snowfall and warmer temperatures,
caribou were probably able to recolonize the large
region in the southern Yukon as a result of their ability
to expand in numbers and migrate into newly available
habitats as cooler temperatures of Little Ice Age pre-
vailed (Viau et al. 2006; Bunbury & Gajewski 2009).
In the face of current climate change, with temperature
increases predicted to be an order of magnitude larger
than that observed during the Holocene (Viau et al.
2006), preserving connectivity between isolated patches
of caribou habitat may become more important than the
individual preservation of small, isolated sedentary
herds. With ongoing research into the recoverability and
usability of STR markers from ancient samples, we hope
to further test these hypotheses.
Acknowledgements
We thank G. Zazula (Yukon Palaeontology, YTG), G. Hare
(Yukon Archaeology, YTG) and the Parks Canada DNA
Repository for providing samples. We thank M. Hart, D. Yang,
R. Farnell, D. Froese, the SFU evolution group (FAB*) and the
Oxford University aDNA laboratory for discussion and
comments. This work was supported by the Canadian Natural
Sciences and Engineering Research Council (NSERC Canada)
through grants to TSK and AOM, grants to TSK from the
Association of Canadian Universities for Northern Studies, the
Beverly and Qamanirjuaq Caribou Management Board and
from Heritage Branch, YTG, and a MEST-CT-2004-007909
(GeneTime) grant to BS.
References
Anderson E (1984) Who’s who in the Pleistocene: a
mammalian bestiary. In: Quaternary Extinctions: A Prehistoric
1 32 2 T. S . KUHN ET AL.
Revolution (eds Martin PS, Klein RG), pp. 40–84. The
University of Arizona Press, Tucson, Arizona.
Bandelt H-J, Forster P, Rohl A (1999) Median-joining networks
for inferring intraspecific phylogenies. Molecular Biology and
Evolution, 16, 37–48.
Berger J (2004) The last mile: how to sustain long-distance
migrations in mammals. Conservation Biology, 18, 320–331.
Bergerud AT (2000) Caribou. In: Ecology and Management of Large
Mammals in North America (eds Demarais S, Krausman PR),
pp. 658–693. Prentice Hall, Upper Saddle River, New Jersey.
Bishop MD, Kappes SM, Keele JW et al. (1994) A genetic
linkage map for cattle. Genetics, 136, 619–639.
Boertje RD, Gardner CL (2000) The Fortymile caribou herd:
novel proposed management and relevant biology 1992-1997.
Rangifer, Special Issue No. 12, 17–38.
Bostock HS (1952) Geology of northwest Shakwak Valley,
Yukon Territory, Geological Survey of Canada Memoir, 267.
Bunbury J, Gajewski K (2009) Postglacial climates inferred
from a lake at treeline, southwest Yukon Territory, Canada.
Quaternary Science Reviews, 28, 354–369.
Clague JJ, Evans SG, Rampton VN, Woodsworth GJ (1995)
Improved age estimates for the White River and Bridge
River tephras, Western Canada. Canadian Journal of Earth
Sciences, 32, 1172–1179.
Cooper A, Poinar HN (2000) Ancient DNA: do it right or not
at all. Science, 289, 1139–1139.
COSEWIC (2002) COSEWIC assessment and update status
report on the woodland caribou Rangifer tarandus caribou
in Canada. Committee on the Status of Endangered Wildlife,
Canada. p. xi +98.
Drummond AJ, Rambaut A (2007) BEAST: Bayesian
evolutionary analysis by sampling trees. BMC Evolutionary
Biology, 7, 214.
Drummond AJ, Rambaut A, Shapiro B, Pybus OG (2005)
Bayesian coalescent inference of past population dynamics
from molecular sequences. Molecular Biology and Evolution,
22, 1185–1192.
Dueck GS (1998) Genetic Relations and Phylogeography of
Woodland and Barrenground Caribou. University of Alberta,
Edmonton, Alberta.
Dupanloup I, Schneider S, Excoffier L (2002) A simulated
annealing approach to define the genetic structure of
populations. Molecular Ecology, 11, 2571–2581.
Evanno G, Regnaut S, Goudet J (2005) Detecting the number of
clusters of individuals using the software STRUCTURE: a
simulation study. Molecular Ecology, 14, 2611–2620.
Falush D, Stephens M, Pritchard JK (2003) Inference of
population structure: extensions to linked loci and correlated
allele frequencies. Genetics, 164, 1567–1587.
Farnell RS, Gardner CL (2003) Status Report on the Chisana
Caribou Herd – 2002. Report TR-03-01. Yukon Fish and
Wildlife Branch, Whitehorse, Yukon Territory.
Farnell RS, MacDonald J (1988) The Demography of the Yukon
Finlayson Caribou Herd 1982–1987. Interim Report. Yukon Fish
and Wildlife Branch, Whitehorse, Yukon Territory, p. 54.
Farnell RS, MacDonald J (1989) Inventory of the Wolf Lake
Caribou Herd. Final Report. Yukon Department of Renewable
Resources, Yukon Territory, p. 70.
Farnell RS, Russell DE (1984) Wernecke Mountain Caribou
Studies. Final Report. Yukon Fish and Wildlife Branch,
Whitehorse, Yukon Territory.
Farnell RS, Florkiewicz R, Kuzyk G, Egli K (1998) The status of
Rangifer tarandus caribou in Yukon, Canada. Rangifer, Special
Issue No. 10, 131–137.
Farnell RS, Hare PG, Blake E et al. (2004) Multidisciplinary
investigations of alpine ice patches in the Southwest Yukon,
Canada: paleoenvironmental and paleobiological
investigations. Arctic, 57, 247–259.
Flagstad Ø, Røed KH (2003) Refugial origins of reindeer
(Rangifer tarandus L.) inferred from mitochondrial DNA
sequences. Evolution, 57, 247–259.
Fluxus Engineering (2008) Network v.4.5. Available from
http://www.fluxus-engineering.com.
Hare PG, Greer S, Gotthardt RM, Strand D (2005) The effect of
the White River Ash on the archaeological record - a view
from the Yukon alpine.
Hasegawa M, Kishino K, Yano T (1985) Dating the human-ape
splitting by a molecular clock of mitochondrial DNA. Journal
of Molecular Evolution, 22, 160–174.
Hayes RD, Farnell RS, Ward RMP et al. (2003) Experimental
reduction of wolves in the Yukon: ungulate responses and
management implications. Wildlife Monographs, (152), 1–35.
Hebblewhite M, Whittington J, Bradley M et al. (2007)
Conditions for caribou persistence in the wolf-elk-caribou
systems of the Canadian Rockies. Rangifer, Special Issue No.
17, 79–91.
Helama S, Timonen M, Holopainen J et al. (2009) Summer
temperature variations in Lapland during the Medieval
Warm Period and the LIttle Ice Age relative to natural
instability of thermohaline circulation on multi-decadal and
multi-centennial scales. Journal of Quaternary Science, 24, 450–
456.
Hinkes MT, Collins GH, Van Daele LJ et al. (2005) Influence of
population growth on caribou herd identity, calving ground
fidelity, and behavior. Journal of Wildlife Management, 69,
1147–1162.
Jagger TA (1945) Volcanoes declare war. Logistics and Strategy
of Pacific Volcano Science. Paradise of the Pacific Ltd,
Honolulu, USA.
James ARC, Stuart-Smith AK (2000) Distribution of caribou
and wolves in relation to linear corridors. Journal of Wildlife
Management, 64, 154–159.
James ARC, Boutin S, Hebert DM, Rippin AB (2004) Spatial
separation of caribou from moose and its relation to predation
by wolves. Journal of Wildlife Management, 68, 799–809.
Kurten B, Anderson E (1980) Pleistocene Mammals of North
America. Columbia University Press, New York.
Kuzyk GW, Russell DE, Farnell RS et al. (1999) In pursuit of
prehistoric caribou on Thandlat, Southern Yukon. Arctic, 52,
214–219.
Lakeman TR, Clague JJ, Menounos B et al. (2008) Holocene
tephras in lake cores from northern British Columbia,
Canada. Canadian Journal of Earth Sciences, 45, 935–947.
Leonard JA (2008) Ancient DNA applications for wildlife
conservation. Molecular Ecology, 17, 4186–4196.
Lerbekmo JF (2008) The White River Ash: largest Holocene
Plinian tephra. Canadian Journal of Earth Sciences, 45, 693–
700.
McDevitt AD, Mariani S, Hebblewhite M et al. (2009) Survival
in the Rockies of an endangered hybrid swarm from
diverged caribou (Rangifer tarandus) lineages. Molecular
Ecology, 18, 665–679.
� 2010 Blackwell Publishing Ltd
ADNA REVEALS RECENT PARTIAL REPLACEMENT 1 32 3
McLoughlin PD, Dzus E, Wynes B, Boutin S (2003) Declines in
populations of woodland caribou. Journal of Wildlife
Management, 67, 755–761.
Malde HE (1964) The ecological significance of some
unfamiliar geological processes. In: The reconstruction of past
environments. Proceedings of the Fort Burgwin Research Center
on Paleoecology 1962. Fort Burgwin Research Center, Fort
Burgwin, USA.
Moodie DW, Catchpole AJW, Abel K (1992) Northern
Athapaskan oral traditions and the White River volcano.
Ethnohistory, 39, 169–180.
Polzin T, Daneschmand SV (2003) On Steiner trees and
minimum spanning trees in hypergraphs. Operations Research
Letters, 31, 12–20.
Post E, Forchammer MC (2008) Climate change reduces
reproductive success of an Arctic herbivore through trophic
mismatch. Philosophical Transactions of the Royal Society of
London, B, 363, 817–818.
Pritchard JK, Stephens M, Donnelly P (2000) Inference of
population structure using multilocus genotype data.
Genetics, 155, 945–959.
Ramakrishran U, Hadly EA, Mountain JL (2005) Detecting past
population bottlenecks using temporal genetic data.
Molecular Ecology, 14, 2915–2922.
Rambaut A (1996) Se-Al: Sequence Alignment Editor, v2.0a11.
Available from http://tree.bio.ed.ac.uk/software/seal/.
Rambaut A, Drummond AJ (2007) Tracer v1.4. Available from
http://beast.bio.ed.ac.uk/Tracer.
Robinson SD (2001) Extending the Late Holocene White River
Ash distribution, Northwestern Canada. Arctic, 54, 157–161.
Schaefer JA (2003) Long-term range recession and the
persistence of caribou in the Taiga. Conservation Biology, 17,
1435–1439.
Shapiro B, Drummond AJ, Rambaut A et al. (2004) Rise and
Fall of the Beringian Steppe Bison. Science, 306, 1561–1565.
Southern Lakes caribou recovery program (1996) Progress
Report: 1992–1996. Yukon Fish and Wildlife Branch,
Whitehorse, Yukon Territory.
Thomas DC, Gray DR (2002) Update COSEWIC status report
on the woodland caribou Rangifer tarandus caribou in Canada.
In: COSEWIC assessment and update status report on the
Woodland Caribou Rangifer tarandus caribou in Canada (eds
Thomas DC, Gray DR), pp. 1–98. Committee on the Status of
Endangered Wildlife in Canada, Ottawa.
Viau AE, Gajewski K, Sawada MC, Fines P (2006) Millennial-
scale temperature variations in North America during the
Holocene. Journal of Geophysical Research - Atmosphere, 111,
D09102.
Vors LS, Boyce MS (2009) Global declines of caribou and
reindeer. Global Change Biology, 15, 2626–2633.
Webber MG, Flanigan MD (1997) Canadian boreal forest
ecosystem structure and function in a changing climate:
impact on fire regimes. Environmental Reviews, 5, 145–166.
� 2010 Blackwell Publishing Ltd
Weladji RB, Holand Ø, Almøy T (2003) Use of climatic data to
assess the effect of insect harassment on the autumn weight
of reindeer (Rangifer tarandus) calves. Journal of Zoology, 260,
79–85.
Wilson GA, Strobeck C, Coffin JW (1997) Characterization of
microsatellite loci in caribou, Rangifer tarandus, and their use
in other artiodactyls. Molecular Ecology, 6, 667–699.
Wittmer HU, Sinclair ARE, McLellan BN (2005) The role of
predation in the decline and extirpation of woodland
caribou. Oecologia, 144, 257–267.
Workman WB (1979) The significance of volcanism in the
prehistory of subarctic North America. In: Volcanic Activity
and Human Ecology (eds Sheets PD, Grayson DK), pp. 339–
371. Academic Press, New York.
Yukon Environment Geomatics (2008) Caribou Herd Ranges
2008, Yukon Territory. Department of Environment, Yukon
Territorial Government, Whitehorse, Yukon.
Zittlau KA (2004) Population Genetic Analyses of North American
Caribou (Rangifer Tarandus). University of Alberta, Edmonton,
Alberta.
Zittlau KA, Coffin J, Farnell RS, Kuzyk GW, Strobeck C (2000)
Genetic relationships of three Yukon caribou herds deter-
mined by DNA typing. Rangifer, Special Issue No. 12, 59–62.
T.S.K. is completing his MSc at SFU under the supervision of
A.Ø.M., with an interest centred on the belief that palaeontol-
ogy can provide crucial information needed to address current
conservation problems. A.Ø.M.’s work concerns explaining
patterns of biodiversity, and in applying phylogenetic data to
conservation problems. K.A.M. is a conservation geneticist
interested in improving management of large ungulates in
North America. P.G.’s research focuses on Arctic ungulate
evolution, in particular through the use of aDNA analysis. B.S.
is an aDNA specialist whose broad research interests include
a focus on the phylogeography and population genetics of
Beringian megafauna.
Supporting Information
Additional supporting information may be found in the online
version of this article.
Table S1 mtDNA primer information
Appendix S1 Sample information.
Appendix S2 Detailed mtDNA methodology.
Please note: Wiley-Blackwell are not responsible for the content
or functionality of any supporting information supplied by the
authors. Any queries (other than missing material) should be
directed to the corresponding author for the article.
Top Related